Agent Autonomy Levels

Sets autonomy levels L0-L3 per task for an AI agent from its track record (review volume, edit and reject rates, breaches), with promotion evidence, automatic demotion triggers and fixed guardrails. Use when asking whether an agent can send unattended, if we can stop reviewing every draft, how much to automate, or what to do after an agent mistake. Not for staging outreach (use outreach-execution).

The-Utopia-Studio Updated 2 repo stars

File contents

the-utopia-studio/skills commit 2e1542529a

Frequently asked questions

npx get-skill add the-utopia-studio/agent-autonomy-levels